Office of Operations
21st Century Operations Using 21st Century Technologies

City of San Francisco, California: Advanced Transportation & Congestion Management Technologies Deployment Initiative (ATCMTD)- 2017

Section II: System Elements

System Elements

The Next Generation Transit Customer Information System is a complex initiative requiring integration among multiple new and existing systems. To manage the project, the SFMTA is proposing to divide it into five primary modules as shown in Figure : (1) Intelligent Predictions Software, (2) Stationary Digital Signage, (3) On-Board Digital Signage, (4) a Mobile Platform and Website, and (5) an Analytics Platform.

Figure 6: System Elements

System Elements

Illustrating the complexity and interconnectedness of this project, Figure 7 is a schematic example of how the Next Generation Transit Customer Information System would integrate with other major SFMTA software and hardware systems. Below are technical descriptions of each system element.

Figure 7: Systems Integration with Other Hardware and Software

Figure : Systems Integration with Other Hardware and Software

Intelligent Predictions Software

The heart of the next generation system is its Intelligent Predictions Software. The fundamental purpose of this element is to generate vehicle arrival and load predictions for downline stops. To accomplish this, it would take inputs from the following systems:

Computer Aided Dispatch/Automatic Vehicle Location (CAD/AVL) System

The CAD/AVL system tracks Muni vehicles, their locations and their corresponding operators, route and schedule assignments. It reports vehicle locations every 60 seconds through a radio communications system.

Automatic Train Control System (ATCS)

ATCS, which controls light rail vehicles on the underground Muni Metro system, monitors the positions of trains in real time. This information is then provided to the CAD/AVL system.

Automatic Passenger Counters (APC)

Starting in 2015, the SFMTA began equipping all rubber-tire vehicles (trolley and motor coaches) and light rail vehicles with state-of-the-art Automatic Passenger Counter (APC) sensors that observe ridership and calculate occupancy in real-time. The CAD/AVL system processes this data and delivers it to SFMTA databases every 60 seconds. The project would also include an option of installing APCs on motor coaches and light rail vehicles procured prior to 2015 that the SFMTA anticipates would still be operating through 2025.

Base Functionality

Generate more accurate vehicle arrival predictions by improving forecasting algorithms, which may include machine learning and incorporation of variables like traffic congestion, ridership loads and weather conditions
Provide predictions based on fixed time- point-based schedules, headway-based scheduling, special event/real-time route adjustments and demand-responsive routing
Eliminate "ghost bus" issues where the existing system shows a prediction for a vehicle that never arrives
Show transfer connections (both within Muni and to regional systems like BART and Caltrain), including estimation of connection times
Incorporate information about real-time reroutes and delays from the Transportation Management Center (TMC)
Push other public messages (such as fare and service changes, elevator alerts) to stationary signage, on-board signage and mobile devices
Provide predictions consistent with SFMTA's Transit Signal Priority (TSP) system, which modifies traffic signals to enable rubber-tire vehicles to catch up to schedule
Provide predictions based on fixed time- point-based schedules, headway-based scheduling, special event/real-time route adjustments and demand-responsive routing
Suggest alternative parallel transit routes within walking distance when there are long wait times, service delays and/or overcrowding
Facilitate usage of complementary sustainable transportation options, such as bike sharing and taxis/on-demand transportation services
Forecast potential vehicle overcrowding

Alternatives

One of the most critical "value added" features of the Intelligent Predictions Software is to automatically generate alternatives to push out to stationary digital signage, on-board digital signs when there may be a long wait, service delay or crowding. As illustrated in Figure , in many parts of Muni's comprehensive network there are nearby parallel services. As demonstrated in SFMTA's public outreach survey, informing someone waiting at a stop of a nearby alternative arriving sooner could increase the chances of remaining with Muni from 44% to 82% under the right circumstances. Over time, that customer may be more likely to view Muni system as more reliable and less likely to switch to less sustainable modes.

Figure 8: Muni Network Density and Alternatives

Figure : Muni Network Density and Alternatives.  A street map of the Muni Network with many blue and red lines running through the streets.

In many parts of San Francisco, the density and interconnectedness of the Muni network provides customers with multiple paths to reach their destination. Taking advantage of the robustness of this network, the new Customer Information System aims to display different alternatives if the initial choice is subject to a long wait, service delay or overcrowding.

Stationary Digital Signage

Figure : Existing Shelter Locations with Digital Signage.  Green - Shelter with Digital Signage.  Green is predominant and is most intense in the Northern part of the San Francisco.  Blue - Other Stops.  It too is spread throughout the city with a higher concentration in the southern part of the city.

Figure 9: Existing Shelter Locations with Digital Signage

Although San Francisco is a hub of the tech industry, many customers do not own smartphones or maintain a data plan. In addition, the City welcomes millions of visitors each year who, even with a smartphone, may not have a data plan and may need additional tools to navigate San Francisco's geography and complex public transportation system. To ensure equitable access to real-time information, the project will include a widespread signage network so that customers have access to real-time arrivals and system status without needing to rely on a smartphone.

Signage at Powered Shelters

The new system would replace the over 850 current Light Emitting Diode (LED) signs at shelters and rail platforms with Liquid Crystal Display (LCD) (or similar technology) signs with LTE communications and WiFi capability. Unlike the current LED signs that show text-only, the new signs would enable a more flexible and user-friendly graphical interface that allows for graphics and text in multiple languages.

Signage at Unpowered Shelters

San Francisco currently has approximately 150 shelters without signs. Providing power to these locations has proven to be technically infeasible and/or cost prohibitive. The SFMTA is reaching out to vendors to explore the potential for cost-effective, solar-powered sign solutions for these locations.

Figure : Potential Solar-Powered Digital Signage

Figure 10: Potential Solar-Powered Digital Signage

(Photo does not imply SFMTA endorsement of a particular vendor).

Signage at Stops without Shelters

There are also over 2,500 stops without a shelter, some of which may be suitable for real-time information signs. Stops without shelters are primarily in lower ridership areas, which often have less frequent service. Therefore, it is all the more important to have real-time signage to inform customers of their expected wait. The SFMTA is seeking cost-effective, solar-powered sign solutions for these locations.

Informational Kiosks

The project would also include a limited number of informational kiosks at locations other than transit stops such as areas with high visitor and tourist traffic. They would assist in pedestrian wayfinding and provide directions to nearby transit stops, in addition to showing real-time transit arrival predictions.

On-Board Digital Signage

In the United States, digital signage on board vehicles is currently limited to displaying and announcing the next stop or next few stops, showing pre-recorded messages, and perhaps listing transfer routes whether or not they are operating. The SFMTA would like to take advantage of new technology to do much more. Examples of new information include:

Transfer Opportunities and Connection Times

Following international best practices, on-board signage with real-time transfer connections (Figure ) will notify customers when their transfer point is approaching and help them better manage the "last mile" leg of their trip. While connecting services may appear physically on a system map, not all routes operate at all times. Real-time information on-board vehicles can tell a customer whether their connecting bus is just a few minutes away or not operating at all, allowing them to assess their options and avoid attempting to make a non-existent connection.

Service Delays and Disruptions

The SFMTA would like to integrate its Computer-Aided Dispatch/Automatic Vehicle Location (CAD/AVL) and radio systems with on-board signage to enable to send timely messages about line reroutes, delays and cancellations from its Transportation Management Center to the on-board digital displays. This will enable customers to keep continually informed about service status and let them decide whether to exit the vehicle if there is a service delay or disruption.

The SFMTA would like to install digital signage on-board approximately 842 buses (New Flyer electric trolley coaches and motor coaches). The 215 next generation Siemens Light Rail Vehicles currently in delivery or on order come with on-board screens. These LCD screens announce stops based on GPS location and pre-recorded messages under a content management system that allows the SFMTA to upload changes via WiFi when a vehicle arrives at a division. Under the Next Generation Transit Customer Information System project, back-end systems and communications would need to be upgraded in order to provide real-time information to Siemens Light Rail Vehicle screens.

Figure : On-Board Digital Signage showing arrival times.

Figure 11 : On-Board Digital Signage

Following international best practices, SFMTA on-board signage could alert customers of intersecting routes and their connection times to reduce the uncertainty and trepidation associated with transfers. (Note: Photo does not imply SFMTA endorsement of a particular vendor.)

Mobile Platform and Website

image of phone with Muni app displayed

Figure 12: MuniMobile Platform Trip

planning and real-time information will be available through an all-in-one mobile app accessible from the MuniMobile platform.

This project will convey customer information in a variety of online formats, including a standard desktop website, a tablet, and a mobile app. Given the widespread use of smart- phones in particular, it is important to deliver accurate and timely information through a mobile app. SFMTA's public outreach efforts have revealed that customers often rely on multiple mobile apps - sometimes two, three or even more - for transit information. This suggests that even the most popular apps do not have all the necessary features that customers are seeking to plan their trips.

The SFMTA now has a separate account-based mobile platform called MuniMobile, which currently offers mobile ticketing and a link to predictions generated by the current real-time information provider. The future system will include a customized app with the capability to integrate with any MuniMobile platform

(Figure). The mobile app will enable customers to plan their trips, follow along while they are enroute to their destination, rate their travel experience and provide feedback and requests relating to service. It will also feature information about sustainable transportation options such as bike sharing and partner transit services.

Rather than create a mobile app from scratch, the SFMTA is open to partnering with an established company that it would designate and endorse as its official and preferred mobile app. However, in contrast to relying on third-party developers to produce independent mobile apps, a widespread model in the United States, the SFMTA will manage app content to ensure that directions and predictions are subject to quality assurance standards.

The app will also have the capability of supporting foreign languages such as Chinese and Spanish commonly spoken by San Franciscans.

The proposed model will also enable the SFMTA to better understand how people use the San Francisco's transportation system. As is standard with customer-focused businesses, understanding consumer preferences with emerging technologies enabled by mobile phones is essential to designing user-focused public services. In conformance with federal, state and local laws and regulations as well as industry best practices, and with user consent, the mobile app and website will be capable of collecting basic locational data to provide context-appropriate vehicle predictions and trip planning and to assist SFMTA with service and operational planning.

Analytics Platform

Once implemented, the Next Generation Transit Customer Information System will provide an advanced analytics platform to help the SFMTA uncover user needs and make service and operational planning decisions that are both data-driven and user-centered. Transaction-level data generated by the proposed Intelligent Predictions Software and the Mobile Platform and Website modules will help calibrate ridership models and project resources required to address unmet or underserved transit demand.

The Analytics Platform will store and process transaction-level data to provide customizable reports, dashboards, and other tools. It will include visualizations for both high-level and in-depth analysis, permitting the SFMTA to assess trends and address issues. The SFMTA will also store this data in its own data warehouse to answer ad hoc queries and for archival purposes.

Data Analytics & Interpretation

Traditional real-time vehicle arrival systems have focused on providing accurate predictions to make easier and less time consuming for people to ride transit. With advancements in technology, particularly in mobile devices as well as processing and visualization of big data, the Next Generation Transit Customer Information System will vastly improve the SFMTA's understanding of the transportation system. Transit agencies have made great strides in measuring operational performance. The next major challenge is to understand and quantify how transit information and service quality impact customer travel choices. This knowledge will inform service and operational planning decisions, help shape transportation policy and ultimately increase ridership.

The new system will generate and store a vast amount of data. The SFMTA envisions partnering with an independent third party, possibly an academic institution, to interpret this data and create models that answer pressing questions. Examples of subjects that the analytics platform could help answer include:

Performance Management

On-Time Performance
What are the system's basic on-time performance statistics by route, route segment and time of day?
Travel Time Variation
How do vehicle travel times vary from time period and from day-to-day along different route
Real-time Prediction Accuracy
How reliable are real-time predictions? What algorithm changes will improve prediction accuracy?

Service and Operational Planning

Service Interventions
What real-time service intervention strategies are most effective in minimizing customer inconvenience and delays?
Interval Reliability
Where are bunches and gaps most likely to occur? How do bunches and gaps affect the predictability of end-to-end customer travel times?
Transfer Reliability
During owl periods, do vehicles arrive on time so that customers can make transfers successfully when there are scheduled timed transfers?
Network Connectivity
How do changes in the network affect customers' ability to move around efficiently?

Customer Responsiveness to Service and Operational Reliability

Mode Choice & Abandonment
How often do potential customers look up the next Muni arrival time and either take Muni or use another transit mode? When does this happen?
Wait Tolerance
How long are customers willing to wait for Muni? How does this wait time vary by time of day, route and location?
Ridership Elasticity
How do service changes affect ridership at a route and network level?
Crowding
How much crowding are customers willing to accept before choosing a different transportation mode?
Origin- Destination Patterns
At an aggregated level, what route(s) are customers taking to travel from their origin to their destination? Is their trip linking with other transit providers or other transportation modes?
Latent Demand
Are there many request for next Muni arrival times when service is sparse and ridership is low? If so, this may suggest that latent demand that could materialize with longer service hours and/or more frequent service.
Customer Feedback
How do service and operational reliability issues impact public perceptions in terms of customer ratings, requests and other feedback?

Figure 13 : Summary of Envisioned Next Generation Transit Customer Information System Features and Innovations

Envisioned System Feature Current Future Innovation*
Intelligent Predictions Software
Real-Time Prediction Algorithm Yes Yes
(improved accuracy)
Yes
(more sophistication)
Real-Time Crowding Alerts No Yes Yes
Real-Time Alternative Route Suggestions No Yes Yes

Real-Time Transfer Connections within Muni and with other systems

No Yes Yes
Stationary Digital Signage
Powered Shelters Yes Yes
(LCD or similar)
No Value
Unpowered Shelters & Stops No Yes
(LCD or similar)
Yes
On-Board Digital Signage
Stop Announcements Yes Yes No Value
Real-Time Transfer Connection Times No Yes Yes
Real-Time Service Delay & Reroute Alerts No Yes Yes
Mobile Platform & Website
Mobile App Yes Yes
(Enhanced capabilities)
No Value
Data Collection No Yes Yes
Mobile Platform & Website
Data Interpretation - Transit Operations and Performance Management Yes Yes
(Enhanced capabilities)
No Value
Data Interpretation - Customer Usage and Travel Preferences No Yes Yes

*Not currently in widespread use in the United States [Return to asterisk]

Partnerships with the Private Sector and Public Agencies

The success of the Next Generation Transit Customer Information System will depend on strong partnerships. The SFMTA will be working closely with the following groups:

Community Stakeholders and Advocacy Groups

The SFMTA is listening to how people currently use information and what they would like to see in the new system. The agency is reaching out to the business community, tourism industry and advocacy/public interest groups such as San Francisco Bay Area Planning and Urban Research Association (SPUR) and the San Francisco Transit Riders. The SFMTA is also engaging organizations representing youths (San Francisco Unified School District and the Youth Commission) and seniors and people with disabilities (Senior and Disability Action, Independent Living Resource Center, LightHouse for the Blind and Visually Impaired, and SFMTA's Multimodal Accessibility Advisory Committee (MAAC)).

Other Transit Agencies

Muni connects with other transit systems such as AC Transit, BART, Caltrain, Golden Gate Transit, SamTrans and ferries that link San Francisco with other parts of the Bay Area. The SFMTA will partner with those agencies for technical assistance in incorporating information about connecting services so that customers can transfer seamlessly between systems.

Vendors

This spring, the SFMTA released a Request for Information (RFI) for the Customer Information System to assess the state of customer information in the transit industry. The SFMTA received responses from 24 entities, excluding potential subcontractors, reflecting widespread interest in the industry to innovate in real-time information delivery. We expect to partner with more than one private sector partner on project implementation.

Academic Institutions

Through the procurement process, the SFMTA will be seeking an analytics platform and data interpretation services to assess the project's success and to answer the performance management, service and operational planning, and customer responsiveness questions. Based on the proposals submitted, it is possible that the SFMTA may select an academic or research institution to provide those services. The primary benefit would be that the academic or research institution would provide independent analysis, including of the performance of the real-time information system itself.

Regulatory, Legislative & Institutional Obstacles

The SFMTA does not anticipate any substantial regulatory, legislative or institutional obstacles with the Next Generation Customer Information System. The SFMTA has jurisdiction over the management of the transportation system in San Francisco. Signage and other project hardware components will be installed on SFMTA-owned or managed locations.

Knowledge Transfer

The SFMTA is excited that the Next Generation Transit Customer Information System will bring many new innovations and could transform how real-time information can contribute to ridership growth. As a potential partner with the U.S. Department of Transportation through the ATCMTD Initiative, we also believe that we would have an important responsibility to share and disseminate project findings so that other cities and transit systems can learn from our experience. The SFMTA fully expects that knowledge transfer and idea sharing through forums and conferences organized by entities such as the US Department of Transportation, Transportation Research Board (TRB) and American Public Transportation Association (APTA) will be an important part of the project.

While San Francisco has a unique, complex transit operating environment, it shares commonalities with many other cities across the United States. For example, it has grown quickly in recent years, faces transit funding shortfalls and has experienced a surge in ridesharing services and automobiles competing for limited roadway space. As such, the impacts of the Next Generation Transit Customer Information System on public transportation's overall attractiveness and utility will have implications for many other cities. By engaging vendors and developing innovative system requirements, this project will also help define best practices for real-time information systems for peer agencies and cities nationally

This project is large-scale primarily due to the size and complexity of San Francisco's transit net- work and fleet. Aside from some base elements such as a prediction algorithm and software integration, the SFMTA estimates that costs will be roughly proportional to the size of digital signage networks at stationary locations and on-board vehicles. Thus, the project's elements are scalable to any city or region. What we will learn from implementing this system will be particularly applicable for smaller and less extensive transit systems where service is less frequent and it can be more challenging to attract ridership that prioritizes short wait times.

Office of Operations